1) Field of the Invention
The present invention relates to an image pickup device, an automatic focusing method, an automatic exposure method, an electronic flash control method and a computer program, which have a face detection function for detecting a face of a subject.
2) Description of the Related Art
Recently, with a development of personal computers, use of digital cameras has become popular as image input devices thereof. Particularly, in many cases, photographers who do not have special techniques use digital cameras. Therefore, in many digital cameras, the shutter speed, exposure and focusing are automatically set, matched with a subject, in order to avoid a failure in taking a picture. In many cases, the subject for photographing is generally a human object. However, there has not been proposed a digital camera, which specially has a setting mode peculiar to human objects. It is a main object of the present invention to photograph a human object appropriately.
In order to always take a picture, which is in focus, while manually performing focus adjustment of the subject, the distance of which from the camera changes every moment, skilled camera work technique is required. Therefore, an automatic focusing function has been developed, so that even nonprofessionals can always photograph an image, which is in focus. In an image pickup device, which has a conventional automatic focusing function, it is common to perform automatic focusing, designating one or a plurality of areas in the center of a screen as a ranging area. In the above ranging method, however, when a human object is photographed, it is necessary to perform automatic focusing in a composition in which the human object and the ranging area are overlapped on each other, and after the focus is adjusted, to change the composition to a desired composition and photograph. Hence, there is a problem in that it takes time to photograph, with a human object being in focus.
As a technique for determining exposure with a conventional image pickup device such as a digital camera, there are used, (1) a multifractionated photometric method in which a photometric range in a screen is divided into a plurality of photometric areas, and exposure is determined from the photometric result of each photometric area, (2) a center-weighted photometric method in which optical intensity is measured by giving weight to the center of the screen, (3) a partial photometric method in which optical intensity is measured in only several tens % of the center of the screen, and (4) a spot photometric method in which optical intensity is measured in only several % of the center of the screen.
With the conventional photometric methods of (1) to (4), however, there is a problem in that a human object is not always exposed properly. More specifically, in the multifractionated photometric method (1), since the exposure is controlled so that the whole screen is exposed properly, there is a problem in that a human object in the screen may not be exposed properly. With the center-weighted photometric method (2), the partial photometric method (3) and the spot photometric method (4), a human object is exposed properly, only when the photometric area agrees with the position of the human object.
Further, when a color photograph is to be taken, using an electronic flash device, a so-called red-eye phenomenon may occur, in which human eyes appear red or gold. Such red-eye phenomenon occurs because flash of light of a light emitter of an electronic flashing device, which has passed through the pupil of the eye, is reflected by the retinal portion, and the reflected light is revealed in the film. More specifically, many capillary vessels exist in the retinal portion of the eye, and hemoglobin in the blood is red, and hence the reflected light thereof appears reddish. In order to prevent the red-eye phenomenon, there has been heretofore known a technique in which after light for preventing the red-eye phenomenon is emitted so as to reduce the size of the pupil of a human object, the main light emission is performed.
With a camera having a conventional electronic flash device, however, there are various problems described below. Firstly, there is such a problem that when a close-up of a face is to be shot, a place where the electronic flash light is strongly irradiated, such as forehead, reflects the electronic flash light, thereby causing blanking. Secondly, there is such a problem that since backlight judgment has been carried out by a difference between the brightness in the central portion of the screen and the ambient brightness, when a human object is not in the middle of the screen in the backlight situation, automatic electronic flashing for correcting the backlight is not performed, and it is necessary to switch to a forced emission mode in order to emit electronic flashlight. Thirdly, there is such a problem that in order to prevent the red-eye phenomenon and blanking due to the electronic flash light, a user has to carry out red-eye mode switchover and a weak light emitting operation of the electronic flash.
In Japanese Patent Application Laid-Open No. 2001-51338, there is disclosed a camera in which the direction of the face of a subject can be recognized, and when the face turns to a predetermined direction, the subject recording operation can be performed. This camera has, a field image detection unit which detects image information of a field, a face recognition unit which recognizes the face of a subject, based on the image information from the field image detection unit, a judgment unit which judges whether the face is turning to a predetermined direction based on the information from the face recognition unit, and a control unit which allows to perform the subject recording operation according to the judgment result of the judgment unit.
However, with the image pickup device such as a digital camera, comprising a mode in which the shutter speed, exposure and focusing are automatically set depending on a subject, pushing of a release button to photograph the subject depends on the photographer's intention. Therefore, when a human object is to be photographed, even if the photographer recognizes the human object, the release timing may be shifted, thereby losing the photographing chance. In Japanese Patent Application Laid-Open No. 2001-51338, such a technique is used that the direction of the face is recognized, and when the face turns to a predetermined direction, the subject recording operation is performed. Therefore, if the human object moves or the camera is blurred, the face may not be photographed clearly.
In Japanese Patent Application Laid-Open No. H10-334213, there is disclosed a technique in which a face image of a target human object image is automatically recognized from color image data including the human object image, without being affected by the photographing background and illuminance, and an area corresponding to the size of the face image is cut out. According to this technique, the camera comprises, an image acquiring CPU which acquires a color image including a human object image, a color liquid crystal display section for displaying the acquired color image on a screen, a face recognition color memory which preliminarily stores a plurality of face recognition color data for recognizing a face image of the human object image from the color image, a recognition instructing section which instructs implementation of face image recognition processing, a face image recognizing CPU which compares color data in the entire color image area with the face recognition color data, in response to the instruction of implementation of the image recognition processing, to thereby recognize an area of color data which agrees with any of the face recognition color data as a face image, an area designator which designates an area to be cut out of the recognized face image, a frame forming CPU which forms a frame for the area corresponding to the size of the face image, in response to the designation of the cutout area, and a face image cutout CPU which cuts out the face image of an area enclosed by the frame.
However, when a photograph for certificate is taken, there has been heretofore such a problem that the face in the photographed image approaches left or right, or up or down, or inclines. When the face inclines, an operator has to specify the cutout position manually, watching the image, or specify the position of the face, thereby making the operation troublesome. Further, Japanese Patent Application Laid-Open No. H10-334213 is for automatically recognizing the face image of a target human object image from color image data including the human object image, without being affected by the photographing background and illuminance, and cutting out the area corresponding to the size of the face image. Hence, it is not taken into consideration to correct the balance of the size to be cut out or an inclination of the image.